Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

https://github.com/megvii-research/OccDepth

Maybe the first academic open work on stereo 3D SSC method with vision-only input.
https://github.com/megvii-research/OccDepth

camera-based occupancy semantic-scene-completion stereo-camera

Last synced: 4 months ago
JSON representation

Maybe the first academic open work on stereo 3D SSC method with vision-only input.

Host: GitHub
URL: https://github.com/megvii-research/OccDepth
Owner: megvii-research
License: apache-2.0
Created: 2023-02-15T04:41:59.000Z (over 1 year ago)
Default Branch: main
Last Pushed: 2023-04-11T06:17:59.000Z (about 1 year ago)
Last Synced: 2024-01-16T05:31:09.728Z (5 months ago)
Topics: camera-based, occupancy, semantic-scene-completion, stereo-camera
Language: Python
Homepage:
Size: 65 MB
Stars: 250
Watchers: 17
Forks: 20
Open Issues: 9
Metadata Files:
- Readme: README.md
- License: LICENSE

Lists

Awesome-Autonomous-Driving - OccDepth - Aware Method for 3D Semantic Scene Completion. (3. Perception / 3.1.1 Vision based)
awesome-stars - megvii-research/OccDepth - Maybe the first academic open work on stereo 3D SSC method with vision-only input. (Python)

README

        # OccDepth: A Depth-aware Method for 3D Semantic Occupancy Network 

![](https://img.shields.io/badge/Ranked%20%231-Camera--Only%203D%20Semantic%20Scene%20Completion%20on%20SemanticKITTI-blue "")

[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/occdepth-a-depth-aware-method-for-3d-semantic/3d-semantic-scene-completion-on-semantickitti)](https://paperswithcode.com/sota/3d-semantic-scene-completion-on-semantickitti?p=occdepth-a-depth-aware-method-for-3d-semantic)

	

[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/occdepth-a-depth-aware-method-for-3d-semantic/3d-semantic-scene-completion-on-nyuv2)](https://paperswithcode.com/sota/3d-semantic-scene-completion-on-nyuv2?p=occdepth-a-depth-aware-method-for-3d-semantic)

# News

- **2023/03/30** Release trained models on GeForce RTX 2080 Ti.

- **2023/02/28** Initial code release. Both Stereo images and RGB-D images inputs are supported.

- **2023/02/28** Paper released on [Arxiv](https://arxiv.org/abs/2302.13540).

- **2023/02/17** Demo release.

# Abstract

In this paper, we propose the first stereo SSC method named OccDepth, which fully exploits implicit depth information from stereo images (or RGBD images) to help the recovery of 3D geometric structures. The Stereo Soft Feature Assignment (Stereo-SFA) module is proposed to better fuse 3D depth-aware features by implicitly learning the correlation between stereo images. In particular, when the input are RGBD image, a virtual stereo images can be generated through original RGB image and depth map. Besides, the Occupancy Aware Depth (OAD) module is used to obtain geometry-aware 3D features by knowledge distillation using pre-trained depth models.

# Video Demo

Mesh results compared with ground truth on KITTI-08:







Voxel results compared with ground truth on KITTI-08:







Full demo videos can be downloaded via `git lfs pull`, the demo videos are saved as "assets/demo.mp4" and "assets/demo_voxel.mp4". 

# Results

## Trained models

The trained models on GeForce RTX 2080 Ti are provided:

| Config| dataset |IoU| mIoU |  Download |

| :---: | :---: | :---: | :---: | :---:|

| [config](occdepth/config/semantic_kitti/multicam_flospdepth_crp_stereodepth_cascadecls_2080ti.yaml) | SemanticKITTI | 41.60| 12.84|[model](https://drive.google.com/file/d/1MGJ_HZcuW5UpULpOeJV0M5ZrT-98j7OE/view?usp=share_link) |

| [config](occdepth/config/NYU/multicam_flosp_crp_stereodepth_cascadecls_2080ti.yaml) | NYUv2 | 49.23| 29.34|[model](https://drive.google.com/file/d/1tBKB-J6NAxDTRTOE1hwAacRmwu57q8L8/view?usp=share_link)|

Note: If you want to get better results, you should set `share_2d_backbone_gradient = false`, `backbone_2d_name = tf_efficientnet_b7_ns` and `feature = feature_2d_oc = 64 (SemanticKITTI)` which needs more GPU memory.

## Qualitative Results





Fig. 1: RGB based Semantic Scene Completion with/without depth-aware. (a) Our proposed OccDepth method can detect smaller and farther objects. (b) Our proposed OccDepth method complete road better.



## Quantitative results on SemanticKITTI



Table 1. Performance on SemanticKITTI (hidden test set). 

|Method            |Input        | SC  IoU       | SSC mIoU       |

|:----------------:|:----------:|:--------------:|:--------------:|

| **2.5D/3D**      |            |                |                |

| LMSCNet(st)   | OCC        | 33.00          | 5.80           |

| AICNet(st)    | RGB, DEPTH | 32.8           | 6.80           |

| JS3CNet(st)   | PTS        | 39.30          | 9.10           |

| **2D**           |            |                |                |

| MonoScene        | RGB        | 34.16          | 11.08          |

| MonoScene(st) | Stereo RGB | 40.84          | 13.57          |

| OccDepth (ours)  | Stereo RGB | **45.10**      | **15.90**      |



The scene completion (SC IoU) and semantic scene completion (SSC mIoU) are reported for modified baselines (marked with "st") and our OccDepth.

## Detailed results on SemanticKITTI.







## Compared with baselines.







Baselines of 2.5D/3D-input methods. ”∗

” means results are cited from MonoScene. ”/”

means missing results

# Usage

## Environment

1. Create conda environment:

``` bash

conda create -y -n occdepth python=3.7

conda activate occdepth

conda install pytorch==1.13.1 torchvision==0.14.1 torchaudio==0.13.1 pytorch-cuda=11.7 -c pytorch -c nvidia

```

2. Install dependencies:

``` bash

pip install -r requirements.txt

conda install -c bioconda tbb=2020.2

```

## Preparing

### SemanticKITTI

- Download kitti odometry and semantic dataset

    - [SemanticKITTI voxel data (700 MB).](http://www.semantic-kitti.org/assets/data_odometry_voxels.zip)

    - [KITTI Odometry Benchmark RGB images (color, 65 GB) and KITTI Odometry Benchmark calibration data  (calibration files, 1 MB)](https://www.cvlibs.net/datasets/kitti/eval_odometry.php)

- Download preprocessed depth

   - [KITTI_Odometry_Stereo_Depth](https://drive.google.com/file/d/1eJPJ1niczagkJfEv21_RdvYBDUbpaQ0w/view?usp=sharing)

- Preprocessed kitti semantic data

    ``` bash

    cd OccDepth/

    python occdepth/data/semantic_kitti/preprocess.py data_root="/path/to/semantic_kitti" data_preprocess_root="/path/to/kitti/preprocess/folder"

    ```

### NYUv2

- Download NYUv2 dataset

    - [NYUv2](https://www.rocq.inria.fr/rits_files/computer-vision/monoscene/nyu.zip)

- Preprocessed NYUv2 data

    ``` bash

    cd OccDepth/

    python occdepth/data/NYU/preprocess.py data_root="/path/to/NYU/depthbin"

    data_preprocess_root="/path/to/NYU/preprocess/folder"

    ```

### Settings

1. Setting `DATA_LOG`, `DATA_CONFIG` in `env_{dataset}.sh`, examples:

    ``` bash

    ##examples

    export DATA_LOG=$workdir/logdir/semanticKITTI

    export DATA_CONFIG=$workdir/occdepth/config/semantic_kitti/multicam_flospdepth_crp_stereodepth_cascadecls_2080ti.yaml

    ```

2. Setting `data_root`, `data_preprocess_root` and `data_stereo_depth_root` in config file (occdepth/config/xxxx.yaml), examples:

    ``` yaml

    ##examples

    data_root: '/data/dataset/KITTI_Odometry_Semantic'

    data_preprocess_root: '/data/dataset/kitti_semantic_preprocess'

    data_stereo_depth_root: '/data/dataset/KITTI_Odometry_Stereo_Depth'

    ```

## Inference

``` bash

cd OccDepth/

source env_{dataset}.sh

## move the trained model to OccDepth/trained_models/occdepth.ckpt

## 4 gpus and batch size on each gpu is 1

python occdepth/scripts/generate_output.py n_gpus=4 batch_size_per_gpu=1

```

## Evaluation

``` bash

cd OccDepth/

source env_{dataset}.sh

## move the trained model to OccDepth/trained_models/occdepth.ckpt

## 1 gpu and batch size on each gpu is 1

python occdepth/scripts/eval.py n_gpus=1 batch_size_per_gpu=1

```

## Training

``` bash

cd OccDepth/

source env_{dataset}.sh

## 4 gpus and batch size on each gpu is 1

python occdepth/scripts/train.py logdir=${DATA_LOG} n_gpus=4 batch_size_per_gpu=1

```

# License

This repository is released under the Apache 2.0 license as found in the [LICENSE](LICENSE) file.

# Acknowledgements

Our code is based on these excellent open source projects: 

- [MonoScene](https://github.com/astra-vision/MonoScene)

- [SSC](https://github.com/waterljwant/SSC)

- [UVTR](https://github.com/dvlab-research/UVTR)

- [CaDDN](https://github.com/TRAILab/CaDDN)

- [BEVDepth](https://github.com/Megvii-BaseDetection/BEVDepth)

Many thanks to them!

# Related Repos

* https://github.com/wzzheng/TPVFormer

* https://github.com/FANG-MING/occupancy-for-nuscenes

* https://github.com/nvlabs/voxformer

# Citation

If you find this project useful in your research, please consider cite:

```

@article{miao2023occdepth,

Author = {Ruihang Miao and Weizhou Liu and Mingrui Chen and Zheng Gong and Weixin Xu and Chen Hu and Shuchang Zhou},

Title = {OccDepth: A Depth-Aware Method for 3D Semantic Scene Completion},

journal = {arXiv:2302.13540},

Year = {2023},

}

```

# Contact

If you have any questions, feel free to open an issue or contact us at [email protected], [email protected].